TLR9 senses single-stranded (ss) DNA. Compared with the relatively uniform shape of the A form dsRNA molecule, the TLR3 ligand, the molecular signature of ssDNA remains obscure. Agonistic ssDNA contains an unmethylated CpG motif that is common to bacterial and viral DNA, as well as a characteristic of apoptotic host DNA. The mechanisms for TLR9-mediated recognition of agonistic DNA have not been fully clarified. Different models have been proposed, such as the "induced-fit" mechanism where conformational changes in the TLR9-ECD are required to initiate TLR9 signaling upon DNA binding, or the "baseless" assumption where the sugar backbone alone in the DNA determine its interaction with TLR9-ECD. Other factors provide versatility for TLR9-mediated DNA recognition, such as HMGB1, a nuclear DNA-binding protein released from necrotic cells, or DNA-antibodies, or malarial hemozoin. Despite the enormous data of TLR9-mediated DNA recognition, important questions regarding the structural basis of DNA induced-TLR9 signaling remain open. Currently we are focusing on the TLR9-ECD consisting of 25 LRRs together with the N and C-terminal capping domains. Our goals are 1) to determine the structure of TLR9-ECD, 2) to define the DNA-binding site, and 3) to explain the observed versatility and specificity of TLR9-mediated DNA recognition at the molecular level. Our ongoing research could help answer some of these questions including if there is conformational change upon DNA binding, whether the binding is baseless, and how TLR9 dimerization occurs. We are currently working on the project of chimeric TLR9 ectodomain. We designed 13 human TLR9 chimeric constructs. The whole C-terminal cap and some of the C-terminal LRRs from TLR9 were truncated and the C-capping domain from human TLR3 was fused. Constructs X1 has 24 LRR from TLR9 and the LRRs were truncated one by one when the other constructs were made. The last, construct X13 has 12 LRR from TLR9. The TLR9 expression clones were made from the bacuovirus/insect cell expression system by PEL at Frederic. All of the chimeric TLR9 constructs were expressed and secreted although only X7 (LRR18) and X8 (LRR17) were secreted well. We purified three constructs (X7, X8 and X10) by FLAG column, Histrap Column and gel filtration column. Initially, all of the desired proteins were aggregated after gel filtration column. This problem was solved by adding appropriate detergent and monomeric proteins were obtained after size exclusion chromatography. Now we are preparing enough amounts of proteins for crystallization setups. We are also studying human TLR8. TLR8 recognizes ssRNA. We are attempting to express the TLR8 ectodomain in sufficient quantity to enable crystallization experiments. The TLR8 expression clones were made from the bacuovirus/insect cell expression system and different expression vectors were explored. TLR8 was expressed well but it was not secreted alone. TLR8 was secreted quite well by the co-expression of one chaperone protein, Calreticulin. The study shows that TLR8 was cleaved after it was secreted from endoplasmic reticulum and this result agreed with the recent publications of TLR9 truncation. However, TLR8 bound tightly to Calreticulin after the purifications by FLAG column, Histrap Column and gel filtration column. The aggregate of TLR8 and Calreticulin was quite stable even though it was treated by high concentration of Urea or Guanadine HCl. We are now working on using other detergents to separate TLR8 from Calreticulin.